Performance of a reversible heat pump/organic Rankine ... · Performance of a reversible heat pump/organic Rankine cycle unit coupled with a passive house to get a positive energy

Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=tbps20

Download by: [University of Liege] Date: 16 December 2016, At: 00:45

Journal of Building Performance Simulation

ISSN: 1940-1493 (Print) 1940-1507 (Online) Journal homepage: http://www.tandfonline.com/loi/tbps20

Performance of a reversible heat pump/organicRankine cycle unit coupled with a passive house toget a positive energy building

Olivier Dumont, Carolina Carmo, Valentin Fontaine, François Randaxhe,Sylvain Quoilin, Vincent Lemort, Brian Elmegaard & Mads P. Nielsen

To cite this article: Olivier Dumont, Carolina Carmo, Valentin Fontaine, François Randaxhe,Sylvain Quoilin, Vincent Lemort, Brian Elmegaard & Mads P. Nielsen (2016): Performance of areversible heat pump/organic Rankine cycle unit coupled with a passive house to get a positiveenergy building, Journal of Building Performance Simulation

To link to this article: http://dx.doi.org/10.1080/19401493.2016.1265010

Published online: 15 Dec 2016.

Submit your article to this journal

View related articles

View Crossmark data

Journal of Building Performance Simulation, 2016http://dx.doi.org/10.1080/19401493.2016.1265010

Performance of a reversible heat pump/organic Rankine cycle unit coupled with a passive houseto get a positive energy building

Olivier Dumonta∗, Carolina Carmob, Valentin Fontainea, François Randaxhea, Sylvain Quoilina, Vincent Lemorta,Brian Elmegaardc and Mads P. Nielsenb

aThermodynamics and Energetics Laboratory, Chemin des chevreuils, 7 B49, 4000 Liege, Belgium; bDepartment of Energy Technology,Aalborg University, Aalborg, Denmark; cDepartment of Mechanical Engineering, DTU, Lungby, Denmark

(Received 29 December 2015; accepted 22 November 2016 )

This paper presents an innovative technology that can be used to deliver more renewable electricity production than thetotal electrical consumption of a building while covering the heat demand on a yearly basis. The technology concept uses aheat pump (HP), slightly modified to revert its cycle and generate electricity, coupled to a solar thermal collector roof. Thisreversible HP/organic Rankine cycle unit presents three operating modes: direct heating, HP and organic Rankine cycle. Thiswork focuses on describing the dynamic model of the multi-component system followed by a techno-economic analysisof the system under different operational conditions. Sensitivity studies include: building envelope, climate, appliances,lighting and heat demand profiles. It is concluded that the HP/ORC unit can turn a single-family house into a PEB undercertain weather conditions (electrical production of 3012 kWh/year and total electrical consumption of 2318 kWh/year) witha 138.8 m2 solar roof in Denmark.

Keywords: heat pump; organic Rankine cycle; positive energy building; dynamic simulation; annual performance

1. Introduction1.1. Context

By 2020, greenhouse gases emissions must be reducedby 20% as compared to the levels of 1990, accordingto European objectives (20-20-20 objectives) (EuropeanCommission 2012). This goal should be achieved throughan increase in the proportion of renewable energy sourcesfrom 9% to 20% together with a 20% increase in systemenergy efficiency. Households account for 27% of the finalenergy consumption (European Commission 2012) andtherefore can constitute an important part of the solution.Various technologies and concepts are being investigated,developed and implemented in the building sector. NetZero Energy Buildings (Marszal et al. 2011) are expectedto gain a significant importance: by 2019, all new buildingsin the European Union should present a renewable energyproduction higher than their primary energy consumption(European Commission 2013).

Net Zero Energy Buildings and, by extension, positiveenergy buildings (PEBs) will therefore play a major rolein the future. PEBs offer different advantages: relativelyhigh independence from energy prices, lower long-termrunning costs and zero fossil-fuel consumption amongothers. Amongst the different available energy sources,solar energy is pointed as a very interesting choice for

*Corresponding author. Email: [email protected]

PEB because it is free, 100% renewable and available inabundance.

1.2. Concept – the reversible heat pump/organicRankine cycle unit

In this paper, the concept of coupling a reversible heatpump (HP)/organic Rankine cycle (ORC) unit to a passivehouse to get a PEB is investigated (Figure 1). A HP/ORCreversible unit is a HP which is slightly modified to be ableto work as an ORC. This reversible unit coupled to a largesolar thermal roof and a horizontal ground heat exchangerconstitutes a combined system able to provide electricityand heat to the household with passive house character-istics. The system can operate in three modes: the directheating (DH) mode uses the heat collected from the roof tosupply the thermal energy in a water store which suppliesthe floor heating (FH) and domestic hot water (DHW).In case of unfavourable meteorological conditions, the HPmode allows to heat the thermal energy store efficiently. Onthe other hand, in case of solar energy availability, the lat-ter is collected and used to cover the heat demand. Finally,a large quantity of heat is generated on the roof duringmid-season and summer periods. This surplus heat can beconverted into electricity by means of the ORC (Dumont,Quoilin, and Lemort 2015).

© 2016 International Building Performance Simulation Association (IBPSA)

2 O. Dumont et al.

Figure 1. The reversible HP/ORC unit integrated in the house (Dumont, Quoilin, and Lemort 2015).

The first investigation on such a system has been intro-duced in 2011 (Schimpf, Uitz, and Span 2011). A thermo-economical tool was developed but only a small area ofcollector (12 m2) and a vertical ground heat exchangerwas considered. In 2013, the modelling and sizing ofsuch a unit has been investigated. The optimal sizingbased on an existing house in Denmark (300 m long hor-izontal ground heat exchanger, 500 l heat storage and138.8 m2 solar roof) lead to a 5 kWe ORC system (Quoilin,Dumont, and Lemort 2013, 2015). A large solar roof isconsidered based on the existing house in Denmark. Themain disadvantage of a large solar roof is the low solarfraction in summer. However, in this study case, thisproblem is avoided thanks to the electricity productionthrough the ORC system with the surplus heat collected inthe roof.

The theoretical results were promising with an ORCelectrical production seven times higher than the electricalHP annual consumption. A prototype has therefore beenbuilt and successfully tested (Dumont, Quoilin, and Lemort2014, 2015). A cycle efficiency of 4.2% was achieved inORC mode (with condensation and evaporation tempera-ture, respectively of 25°C and 88°C) and a COP of 3.1 wasobtained in HP mode (with condensation and evaporationtemperature, respectively of 61°C and 21°C).

1.3. ScopeThe first part of this paper details the models of: thereversible HP/ORC unit, the passive house, the horizontalground heat exchanger and the flat plate solar roof col-lector. Each sub-model, the global model and the controlstrategy of the system are described in detail in Section 2.

The model is then used to simulate and assess theenergy system performance in typical days along the yearfor this innovative concept (Section 3). Followed by a

study of influence including building envelope, location,heat demand, lighting and appliances profiles is performedbased on annual results (Section 4)

Finally, an economic comparison with a HP and photo-voltaic panels (PV) is made.

2. Modelling methodology2.1. Simulation toolAmong simulation programs, some are dedicated to build-ing performance simulation (IDA ICE, ESP-r, EnergyPlus,TRNSYS, WUFI

®Plus, etc.) while others are more general

(Dymola/Modelica, MATLAB/ Simulink, IDA SE, etc.).Simulation tools like Matlab–Simulink need the model tobe implemented, in a state-space form in which causal rela-tions play an important role. A simulation language basedon an object-oriented approach and physically orientedconnections – Modelica – is chosen as simulation tool tomodel the new system proposed in this work. Recently,Modelica has become more and more used in buildingperformance simulation. The Lawrence Berkeley NationalLaboratory developed a Modelica library called Buildings,that contains a large number of HVAC components anda multi-zone building model (Wetter, Zuo, and Nouidui2011). Also, the RWTH Aachen and UdK Berlin (Nytsch-Geusen and Unger 2009) are developing Modelica librariesfor HVAC-systems and building models. Besides, manymodels for HVAC components and different thermal zonemodels, the RWTH Aachen library offers a database ofmanufacturer’s data for building technology (Muller andBadakhshani 2010).

Before describing each sub-model and the control strat-egy, it is important to note that the dynamic modelling of asystem including several sub-systems does not systemati-cally require each model to be dynamic: components char-acterized by relatively low time constants can be modelled

Journal of Building Performance Simulation 3

as quasi-steady-state, since their fast dynamics are not rele-vant to the overall simulation and can substantially impactthe computational effort. In this case, it was shown previ-ously (Perers 1997; Schnieders 1997; Chow 2003; Fischeret al. 2004; Dumont, Quoilin, and Lemort 2014; Freeman,Hellgardt, and Markides 2015) that the dynamics of thereversible unit can be neglected because of its small inertiacompared to other sub-systems.

2.2. Reversible HP/ORC unitAn experimental investigation has been carried out on theunit in HP and ORC mode over a wide range of con-ditions (Dumont, Quoilin, and Lemort 2015). Based onthe measurements, semi-empirical models have been cal-ibrated for each component (heat exchangers, compressor,pump and pipes). These models are then combined tosimulate the behaviour of the global system. Finally, poly-nomial regressions, fitted on the global validated model,allow to evaluate the outputs of the reversible unit. Theseare presented by the authors in a former paper (Dumont,Quoilin, and Lemort 2014). The T-s diagrams for the HPand the ORC are given in the appendix (Figures A1 andA2). A cycle efficiency of 5.3% is achieved in ORC mode(with condensation and evaporation temperature, respec-tively of 25°C and 88°C) and a COP of 4.21 is obtained inHP mode (with condensation and evaporation temperature,respectively of 61°C and 21°C).

2.3. StorageThe basic type of hot water storage tank in the HP/ORCsystem is shown in Figure A3. It is a typical DHW tank sys-tem installed in single-family houses in Denmark (500 l).The water tank consists of a stainless steel cylinder withtwo built-in spiral heat exchangers (HXs) – one going frommid-height to bottom of the tank and another going frombottom to the top of the tank. The working fluid in theHP/ORC unit is circulated through the mid-height heli-cal heat exchanger, while the cold water from the gridis circulated through the all-through heat exchanger tosupply DHW. In the current work, this stratified sensiblethermal storage is modelled by a one-dimensional finite-volume method comprising 20 isothermal segments withequal volume (Carmo et al. 2015). The model accountsfor heat losses to the environment, internal heat conduc-tion between adjacent cells as well as for internal naturalconvection whenever an internal reversed temperature gra-dient occurs. The dynamic temperature profile of the tankis represented by a set of i ordinary differential equationsthat represent the energy balance of the tank (Equation(1)). The first term is the thermal inertia of the cell.The second term is composed (from left to right) of theenthalpy flow, the thermal exchange with an eventual heat

exchanger, conduction with adjacent cells and ambientlosses.

Ai�xρiCPdTi

dt= m(hex,i − hsu,i) + Ahx,iQhx

+ αAi+1Qi+1 + �Ai−1Qi−1

− Aamb,iU(Ti − Tamb). (1)

In this equation α is 0 if the ith node is the top of thetank and 1 otherwise and β is 0 if the ith node is the bot-tom node and 1 otherwise. This model is validated usingexperimental data under different charging and discharg-ing conditions following prEN12977-3:2008 (CEN 2008).More details can be found in a former work (Carmo et al.2015).

2.4. Solar roofThe solar roof currently installed in the house is a prototypeof aluminium pipes installed on an aluminium absorberplate covered with the Alanod Miorosol coating (InnogieAps 2013). A four millimetre thick glass surface is added toensure the glazing (Figure A4). Commonly, thermal panelsare smaller, but in this case it is more interesting to coverthe whole roof (138.8 m2) because it is an integrated tech-nology that acts as a solar collector and a roof. It avoidsthe necessity of buying a classical roof plus solar collec-tors (Innogie Aps 2013). Furthermore, the excess heat insummer is not wasted and can be converted into electricitythrough the ORC. This large roof size is classical for newbuildings in the countryside of Denmark.

The heat collected from the roof is therefore modelledwith Equation (2) involving the useful solar roof area (A),the outdoor temperature (Tamb), the mean absorber temper-ature (Tm), the overall heat transfer coefficient (Uo) andthe solar irradiance absorbed by a collector per unit area ofabsorber (I ).

Qroof = A(I − Uo(Tm − Tamb)). (2)

The overall heat transfer coefficient (Uo) takes intoaccount the top losses, the edge losses and the back losses.The edge losses are assumed to be zero, since the heattransfer is negligible when the collector area is higher than30 m2 (Duffie and Beckham 2006). The back losses are alsoassumed to be zero due to the 400 mm thick insulation atthe back of the collector. Finally, the top loss coefficientis evaluated using Equation (3) with a maximum error of0.3 W/m2 for mean absorber temperatures below 200°C(Klein 1975).

UT =

⎛⎜⎝ 1

vTm

(Tm−TaN+f

)e + 1hw

⎞⎟⎠

−1

+ σ(Tm + Ta)(T2m + T2

a)

1εp+s·N ·hw

+ 2N+n−1+z·εp

εg− N

. (3)

4 O. Dumont et al.

The different terms composing Equation (3) are detailedin the appendix – Table A1. The dynamic model of thesolar roof finally obtained by combining Equation (2) witha thermal inertia corresponding to 104.6 l of 30% volumeglycol-based water solution (4).

Qroof,inertia = Qroof + Mglycol · CPglycol · dTm

dt. (4)

2.5. Building modelThe model is based directly on the geometry and the con-struction characteristics of the real Danish building. Asimplified lumped parametric model is applied. The root-mean-squared error of a such a model related to innertemperature has been shown to be always lower than 1 K(Masy 2007). The arrangement of the different rooms ofthe building and the composition of the walls are takeninto account. The building is first divided into five zones(dining room and kitchen, main bedroom, bathroom, halland toilet and finally guest bedrooms. See zones defini-tions in Figure 2 and zone characteristics in Table A2) withconstant volume, uniform temperature and conservationof mass and energy in each zone. The walls are mod-elled with two thermal resistances and one heat capacity,

parameters being given in Masy (2007). Four inputs areadded in each zone: lighting, appliances, occupancy and athermal exchange with adjacent zones. Wind pressure andbuoyancy from the air-specific volume difference and ven-tilation are not modelled in order to avoid too high levelof complexity and computational time. Finally, the radiantslab (25 m2) from the buildings library (Wetter et al. 2013)is connected to the only room where it exchanges heat inthe house (zone 1).

2.6. Ground source horizontal heat exchanger (GHX)2.6.1. Description of the case-studyThe ground source horizontal heat exchanger consists ofthree layers layout. The layers are linked in parallel andburied, respectively at 0.50, 1.00 and 1.50 m depth. Eachlayer consists of 24 tubes disposed in a head to tail set-ting. The tubes are made in cross-linked polyethylene andare 22.89 m long with a diameter of 2.6 cm (geometry ofthe GHX – Figure A5). A 30% monoethylene glycol–water mixture is used as the heat transfer fluid. The soil isassumed to be argillaceous with a water content of 10%,which corresponds to an average soil moisture (Bircheret al. 2012).

Figure 2. Division of the house into five zones.

Journal of Building Performance Simulation 5

2.6.2. Description of the modelThe deep earth temperature is set to 10°C.This choiceis made following ground measurements conducted inPotsdam, Germany, (PICIR 2015). The absorbance andemissivity of the soil surface are, respectively, set to 0.55and 0.75. An average wind speed of 4 m s−1 is considered.

A model of the ground source horizontal heatexchanger already exists (using the finite element method)in the TRNSYS simulation language (TESS 2006). Areduced-order model is developed and calibrated basedon the reference finite element model (TESS 2006). Thismodel is designed to be flexible and is valid for differentkinds of pipes geometry and layout.

The model consists in discretizing three layers ofground (Figure 3). The central element in the model is the

Figure 3. Layout of the reduced-order model of the horizontalground heat exchanger.

Figure 4. Water outlet temperature of the ground heat exchangersubmitted to step inputs: Finite-element model (Trnsys) versusreduced-order model (Dymola).

soil central thermal mass which simulates the soil directlysurrounding the GHX pipes. In addition, a surface layerwhich reacts rapidly to climate variations (solar irradia-tion, ambient temperature and sky temperature) is added.Finally, a sub-soil layer presenting slow variations throughthe seasons is modelled and connected to the deep earthtemperature. Each layer is modelled with a central capacityand two resistors. The pipes are modelled with a finite-volume 1D flow model (20 cells) from the Thermocyclelibrary (Quoilin et al. 2014). Finally, two thermal resistorsare added to the pipes to account for the resistance of thetube and, for the resistance of the soil.

2.6.3. Calibration of the reduced-order modelThe reduced-order model described here above is cali-brated with the finite element model as a reference by vari-ation of the two main inputs, which are the ambient temper-ature and the solar irradiation. The GHX model parametersare defined in Table A3 (in the appendix). With theseparameters, results show good agreement between the twomodels. A maximum deviation of 0.5 K is observed for theprediction of the water outlet temperature (Figure 4).

2.7. Global modelFigure 5 presents the flowchart of the global model com-bining the storage, the building, the roof, the reversible unitand the ground heat exchanger. Hourly schedules are asso-ciated with the occupancy, the DHW use, the lighting andappliances in each zone (Georges et al. 2013). The weatherdata used for the outdoor temperature and the solar irra-diance are provided by the DMI – Danish MeteorologicalInstitute – (Wang et al. 2010) in the case of Denmark andby Energy Plus Energy Simulation Software (EnergyPlus2015) for other locations. An adaptive time step is com-puted by the solver, but is not allowed to exceed 900 s. Alow time step induces too much computational time andtoo large output file size, a time step larger than 20 mincould lead to errors larger than 5% (Bouvenot et al. 2015).The typical computational time is 3 h for an annual simula-tion. The consumption of auxiliary pumps (except GHX

Figure 5. Global model and connections between sub-models.

6 O. Dumont et al.

pump) is neglected, they represent less than 2% of theglobal system power consumption.

Some parameters have to be fixed: roof water flow rate,ground heat exchanger water flow rate, and storage waterflow rate and temperature set-points of the storage. Practi-cally, the following values are used for the flow rates basedon real values imposed in the house:

• Roof water flow rate = 0.6 kg s−1,• Ground heat exchanger water flow rate = 1.5 kg s−1,• Storage water flow rate = 0.6 kg s−1

These flow rates should be optimized in future inves-tigations to increase the energy efficiency of the system(Burhenne et al. 2013).

2.8. ControlThe control strategy ensures that the heat demand is cov-ered while electricity is produced with the surplus of heat.For this reason the first control variable used is the hotwater storage tank temperature (the control temperaturepoint is located at mid-height of the tank).

A state diagram control is implemented. The conditionsgoverning the transitions between the three modes (HP,ORC and DH) and the stand-by mode (Bypass) are shownin Figure 6. The Bypass mode means that no HP, ORCor DH is activated, only the FH circuit can be activatedextracting energy from the water store, if necessary, toreach the desired indoor conditions (20°C). The principleis the following: if the storage is too cold (the control tem-perature of the storage is lower than the low-temperature

threshold), the HP mode is activated. If the roof tem-perature is higher than the storage one, the DH mode isused. Finally, the ORC system produces electricity whenthe storage temperature has reached a given high thresh-old and if a stable state can be reached. This means thatthe ORC is only activated once it can produce a certainlevel of power (WORC,min). The WORC,min is used to enable asmooth and efficient operation of the system in ORC mode.When a stable operation of the ORC cannot be guaranteed(WORC < WORC,min), the TES is allowed to go above thehigh-temperature threshold. It should be noted that the HPmode, is using either the roof or the horizontal ground heatexchanger depending on which one is the warmest.

Table 1 summarizes the values of each threshold tem-perature. The threshold values were chosen to avoid chat-tering (too many mode changes) and to maximize theefficiency of the system in Dumont, Quoilin, and Lemort(2014). The number of mode changes is considered highwhen more than one change occurs in a 15-min period.

It should be noted that, although the set-points andthresholds have been optimized, the proposed control strat-egy is still a myopic rule-based control strategy. A truly

Table 1. Values of the temperature thresholds.

Temperature threshold Abbreviation Value

High-temperature threshold ofthe storage

Tsto,h(°C) 50

Low-temperature threshold ofthe storage

Tsto,l(°C) 40

Power threshold of the ORC WORC,min(W) 2000Indoor comfort temperature Tin(°C) 20

Figure 6. State diagram control. Troof is the roof exhaust temperature, Tsto is the storage control temperature (middle height of the tank),Tsto,l is the low-temperature threshold of the storage, Tsto,h is the high-temperature threshold of the storage, WORC,min is the minimumpower to start the ORC system.

Journal of Building Performance Simulation 7

optimal control strategy is difficult to implement becauseof the high number of manipulated variables, the numerousset-points and the non-linearity of the problem. It wouldrequire a predictive nonlinear optimization, based on thenext 24 h of weather forecast, user behaviour and elec-tricity prices. Such approach would avoid, for example,starting the HP when the solar heat will be sufficient tocover the heat demand later in the day.

2.9. Performance criteriaYearly simulations are performed and evaluated throughthe following performance criteria:

• Gross electrical production (Wh): the energy pro-duced by the ORC (or the PV panels if specified)(Wel,prod).

• HP electrical consumption (Wh): the electrical con-sumption of the HP (Wel,HP).

• Gross electrical consumption (Wh): the sum ofappliances, lighting and HP electrical consumption.

• Net electrical production (Wh): the gross electricalproduction minus the gross electrical consumption(Wel,net).

• The total thermal energy production of the unit,includes both thermal production from HP and solarthermal roof collector (Wh) (Qth,prod).

• DH energy (Wh): the total thermal energy gained bymeans of the DH mode (QDH).

• B, Benefits (e): the income benefits evaluated fol-lowing the Danish law (Equation (5)). It does nottake any investment into account. WHP is the elec-trical power consumption of the HP. Wnet is thenet electrical power, i.e. the electrical productionminus the electrical power consumption of lightingand appliances. Pr ∼ 0.28eW−1 h−1 is the retailprice considered when the net electrical power isnegative, Pr,HP is the retail price for the HP only∼ 0.22eW−1 h−1 and Pbb is the buy-back tariff ∼0.17eW−1 h−1 considered when the net electricalpower (Wnet) is positive. Retail and buy-back tariffsare provided by real data from Denmark (Energinet2015).

If Wnet > 0 then B =∫ t

0(Pbb (Wnet) − Pr,HP · WHP) · dt

else B =∫ t

0(Pr (Wnet) − Pr,HP · WHP) · dt.

(5)

• Supply cover factor or self-production rate (γS),which represents the fraction of energy producedby the ORC (or PV) which is used to coverinstantaneous electrical consumption (Equation (6))(Baetens et al. 2012).

γS =∑

min (Wcons., Wprod)∑Wprod

. (6)

• Demand cover factor or self-consumption rate (γd),which represents the fraction of energy consump-tion which has been produced by the ORC (or PV)(Equation (7)).

γD =∑

min (Wcons., Wprod)

Wcons.. (7)

For all the simulations in this paper, the set-points tem-perature of the storage, the set-point in the main room ofthe building and the solar roof are the same.

3. Simulation of typical daysThe system response is presented for three characteristicdays in Denmark: a winter day (day 1), a spring day (day62) and a summer day (day 182). Eight variables are anal-ysed in this section: the storage control temperature (Tsto),the outdoor temperature (Tout), the house ambient temper-ature in zone 1 (Tin), the exhaust roof temperature (Troof),the ground heat exchanger exhaust temperature (TGHX),the heat flow rate for FH (QFH), the heat flow rate forDHW (QDHW), the heat flow rate from the reversible unit orfrom the solar roof (Qth,prod) and the electrical unit powerconsumption ( − )/production ( + ) (Wel).

3.1. Winter – Day 1The behaviour of the system is plotted in Figure 7 for acharacteristic winter day. Slightly after 4., the FH is acti-vated (QFH) in a way to keep the indoor temperature (Tin)

close to 20°C. This leads to a decrease in the control tem-perature of the storage (Tsto) down to the lower temperaturethreshold of 40°C. The HP mode is therefore activatedto raise the control temperature of the storage up to thehigh-temperature threshold of the storage (50°C). Thisphenomenon is observed three times during this day (4.,14 and 20). The heat generated in HP mode is QHP/ORC andcorresponds to an electrical power of Wel. The DH modecannot be activated because of the low temperature of thewater in the roof (Troof). In this case, the system is actingas a classical ground source HP during this representativewinter day.

3.2. Spring – Day 62A typical spring day is depicted in Figure 8. First, around3, the FH starts, decreasing the storage control tempera-ture. Thus, the HP is activated following the same schemeas for the typical winter day. The difference is that around10 h 30 min, the roof exhaust temperature is higher thanthe storage temperature and the system can therefore bene-fit from DH until the next day. In that case, the DH allowsto start the HP mode only once during day 62 to cover theheat demand of the building.

8 O. Dumont et al.

Figure 7. Dynamic simulation of the reversible unit coupled to a passive house for the first day of the year.

Figure 8. Dynamic simulation of the reversible unit coupled to a passive house for the 62nd day of the year.

3.3. Summer – Day 182Figure 9 presents the response of the reversible unit for acharacteristic summer day for the study case in Denmark.Slightly before 8 the DH mode is activated since the rooftemperature becomes higher than the storage temperature.When the storage temperature reaches its maximum value,the ORC mode can be activated to generate electricity. The

electrical production of the ORC is low (compared to thenominal power, 5290 W) due to the high temperature ofthe water in the GHX. Since the heat demand is rathersmall (no FH, only DHW) and the capacity of the storageis hot enough there is no need to heat the thermal energystore. The ORC mode is therefore activated as long as theelectrical production is greater than zero.

Journal of Building Performance Simulation 9

Figure 9. Dynamic simulation of the reversible unit coupled to a passive house for the 182nd day of the year. Mode 1 is ORC, mode 2is direct heating and mode 3 is heat pump.

4. Annual performance

4.1. Reference case

First, before establishing a sensitivity analysis, a basic caseyearly simulation corresponding to the real conditions ofthe house located in Herning, Denmark, is performed. Inthis simulation, there is one thermal storage of 500 l forDHW and FH. Figure 10 presents a comparison of the

electrical ORC production, HP electrical consumption andthermal energy provided by the DH mode for each monthof the year. The HP is running during 5 months of the year,mainly in winter, leading to a total electricity consumptionof 827 kWhe and heat supply of 3082 kWhth. DH is used10 months of the year and produces 1207 kWth, represent-ing 28.1% of the total heat demand of the building duringa year.

Figure 10. Comparison of the heat pump electrical consumption, electrical ORC production and thermal energy provided by the directheating mode for each month of the year in the reference case.

10 O. Dumont et al.

The DH mode is less used in summer months comparedto March and October because the heat demand for FH issignificantly lower. The gross electrical ORC productionis equal to 3012 kWhe, the lighting and appliances con-sumption reaches 1491 kWhe, leading to a net electricalproduction of 694 kWhe on a yearly basis. This demon-strates the ability of the current technology to get a PEBin terms of electricity use. Using Equation (3), the annualrunning costs of the system in aforementioned conditionsare 119e.

4.2. Results of the sensitivity analysis on theperformance of the HP/ORC system

After considering the basic case, it is interesting to com-pare the system behaviour resulting from different climates.A former project (Knight et al. 2010) has shown thatEuropean climate can be divided into five different typi-cal zones. The system is therefore simulated for five citieslocated in each zone (from north to south): Copenhagen,Frankfurt, Torino, Rome and Palermo. For comparisonpurposes, the feed-in tariffs for the different locations weremaintained as in the Danish case.

Secondly, two additional different building envelopecharacteristics − K15 and K30 (Masy et al. 2015) are

studied in all climates. They differ in terms of coeffi-cient of heat transmission and air tightness (see appendix– Table A4). Finally – as proposed in Georges et al.(2013) – two additional Light and Appliances profiles(L&A) are simulated with the reference Danish buildingcharacteristics. The latter differ in the magnitude of powerdemand. In descending order of magnitude, L&A 2010(3000 kWh/year) is characterized by highest demand, fol-lowed by L&A 2030 (2000 kWh/ year) and L&A Danish(1491 kWh/year). Table 2 shows the results of sensitiv-ity analysis on the performance of the HP/ORC systemunder different conditions of climate, insulation and lightsand appliances demand according to the performance para-meters listed in Section 2.9 Performance criteria.

From Table 2, it can be concluded that for any build-ing and light and appliances demand sunniest locations(south Europe) leads to higher power production and thus,higher financial benefits. On the other hand, the HP isalmost never used in southern locations, because the heatdemand is small and, therefore can benefit from the DH.On the contrary, northernmost locations present low heatenergy provided by DH. There is an optimal location inlatitude close to Torino that shows the best compromiseto benefit optimally of the thermal energy from the DH. Itis interesting to note that an increase in lights and appli-ances demand – in all locations – not only decreases the

Table 2. Results of the sensitivity analysis.

Building L&A LocationQth,prod(kWh)

Wel,prod(kWh)

Wel,hp(kWh) Wnet(kWh) Benefits (e) γ s γ D Qdh(kWh)

Danish 2010 Copenhagen 3597 3015 690 − 675 − 501 0.13 0.13 1057Frankfurt 3291 3609 572 37 − 368 0.119 0.14 1180Torino 2243 5379 189 2190 38 0.1 0.18 1523Roma 1072 6646 16 3630 312 0.1 0.23 990Palermo 861 8597 0 5597 666 0.096 0.27 845

Danish Copenhagen 4289 3012 827 694 − 119 0.071 0.138 1207Frankfurt 3879 3607 699 1417 15 0.065 0.15 1292Torino 2700 5371 251 3629 422 0.054 0.185 1726Roma 1301 6639 35 5113 695 0.053 0.226 1148Palermo 889 8597 0 7106 1046 0.049 0.27 872

2030 Copenhagen 4025 3014 783 231 − 260 0.093 0.133 1134Frankfurt 3652 3609 647 962 − 125 0.084 0.145 1254Torino 2545 5374 226 3148 281 0.071 0.181 1671Roma 1211 6643 26 4617 553 0.07 0.22 1088Palermo 875 8596 0 6596 904 0.065 0.26 859

K15 Danish Copenhagen 2887 3021 535 995 − 38 0.047 0.096 912Frankfurt 2685 3615 447 1677 81 0.042 0.1 1034Torino 1772 5386 120 3775 464 0.036 0.131 1304Roma 980 6648 12 5145 708 0.036 0.159 917Palermo 863 8596 0 7105 1048 0.033 0.19 847

K30 Danish Copenhagen 8667 2987 1723 − 227 − 318 0.046 0.092 2031Frankfurt 7804 3573 1457 625 − 156 0.041 0.098 2206Torino 5956 5334 803 3040 300 0.035 0.128 2837Roma 3254 6585 196 4898 655 0.035 0.155 2468Palermo 1670 8562 15 7056 1038 0.032 0.187 1586

Notes: (Qth,prod is the total thermal energy production of the HP/ORC unit, in both HP and DH mode, Wel,prod is the gross electricalproduction, Wel,HP is the HP electrical consumption, Wel,net is the net electrical production). B are the income benefits, γ S is the self-production rate, γ d is the self-consumption rate and QDH is the DH energy.

Journal of Building Performance Simulation 11

net power output and benefits, but also decreases the HPpower consumption. This is due to the internal heat gainsby means of light and appliances, which decrease the heat-ing demand. On the other hand, it is shown that lowerlevels of insulation lead to higher heating demand coveredby DH without compromising the ORC power output andthe financial benefits.

4.3. Comparison with a HP combined withphotovoltaic panels

In a former article (Dumont et al. 2015), a perfor-mance comparison between the HP/ORC reversible unitand a classical mature alternative solution for PEBswhich consists of photovoltaic panels combined with awater-to-water heat pump (HP/PV) is performed. Anotheralternative single-technology capable of delivering heatand electric power is PVT but it is considered out of thescope of this study (He et al. 2006; Herrando, Markides,and Hellgardt 2014 and Dupeyrat, Ménézo, and Fortuin2014). In this former paper (Dumont et al. 2015), the areaof photovoltaic panel is fixed in a way that the electricalpeak power is the same as the HP/ORC reversible unitin typical nominal summer conditions to get comparableresults. It is shown that, the best system is always theHP/PV system in terms of electrical production, incomebenefits and matching of the production and consumption.Nevertheless, an interesting advantage of the reversibleunit is the lower HP electrical consumption which makesthis system more profitable if no electricity can be boughtfrom the grid (isolated network for example). Furthermore,an economic feasibility study of the total cost (income ben-efit and investment) of the HP/ORC system is comparedto the cost of the HP/PV system. The reversible system isnever profitable in the base case, i.e. with a heat demandcorresponding to the real house. But, if the heat demandis significantly higher (8 times higher DHW consumption)the reversible unit is much more profitable.

5. ConclusionThe recent interest for PEBs has led to development ofnew technologies and solutions. In this paper, a reversibleHP/ORC coupled to a passive house is studied. This tech-nology is a promising option to achieve a PEB. The mod-elling of each sub-system (ground heat exchanger, thermalenergy storage, building, solar roof, reversible HP/ORCunit) and the control strategy are described extensively.Simulations show that this technology leads to a PEB onan annual basis. Moreover, a sensitivity study has led tothe following conclusions:

• The HP/ORC system presents a positive net electri-cal production while covering the total heat demandof the building over a year, even in cold climatessuch as that of Denmark. The results show that,

in the Danish case-study of a single-family housewith a 138.8 m2 solar collector, the electrical pro-duction by the ORC system yields 3012 kWh/yearwhile the total annual electrical consumption is2318 kWh/year.

• The climate in southernmost cities is much morefavourable for the ORC system because it works forlonger periods and closer to its nominal conditions.

• There is an optimum location (for latitudes aroundTorino) where the DH is maximum.

• A low insulation of the building and/or a low energylighting and appliances profile leads to a better use ofthe system, benefiting from more energy from DH.

• When compared to a HP coupled with PV panels,results show that the HP/ORC unit could only beprofitable in the case of a large heat demand of thebuilding and/or restriction on buying electricity fromthe grid. More generally, this means that buildingswith a high heat demand, everything else being con-stant, are profitable for the reversible unit. A largebuilding or a building with high DHW consump-tion could fit this constraint (office building, hospital,prison, stadium, etc.).

AcknowledgementsThis work was supported and funded by the Eurostars Program(Single HPA Unit).

NomenclatureVariables

A area (m2)B income benefits (e)COP coefficient of performance ( − )CP specific heat capacity at constant pressure (J/(kg K))e empirical variable used in the roof model ( − )f factor ( − )h specific enthalpy (J/(K kg))i index ( − )I irradiance (W m−2)M mass (kg)m mass flow rate (kg/s)N number of plates ( − )P cost (e.W−1 h−1)Q heat transfer (W)r interest rate (%)s empirical constant ( − )t time (s)T temperature (°C)U heat transfer coefficient (W m−2 K−1)v empirical constant used in the roof model ( − )W energy (W h)W Power (W)x length (m)y empirical constant ( − )z empirical constant ( − )

Greek symbols

η efficiency ( − )α numeric coefficient ( − )

12 O. Dumont et al.

� numeric coefficient ( − )β collector tilt (°)� difference ( − )γ cover factor ( − )ζ emittance ( − )ρ density (kg/m3)

Abbreviations

CHP combined heat and powerGHX Horizontal ground heat exchangerHP/ORC Reversible HP/ORC unitHP/PV HP combined with PVNZEB Net Zero Energy BuildingPV Photovoltaic panels

Subscripts

amb ambientb backbb buy-backBH Boreholecons consumptionD demandex exhaustFH (floor) floor heatingglycol GlycolHGHE horizontal ground heat exchangerh highhx exchangerin indoorInertia Inertial lowl-a lighting and appliancesm meanmatch matchingmin minimumnet Netp plateprod productionO Overallout outdoorr retailroof solar roofS supplysto storagesu supplyT top

ReferencesBaetens, R., R. De Coninck, J. Van Roy, B. Verbruggen, J.

Driense, L. Helsen, and D. Saelens. 2012. “AssessingElectrical Bottlenecks at Feeder Level for Residential NetZero-Buildings by Integrated System Simulation.” AppliedEnergy 96: 74–83.

Bircher, S., N. Skou, K. H. Jensen, J. P. Walker, and L. Ras-mussen. 2012. “A Soil Moisture and Temperature Network

for SMOS Validation in Western DK.” Hydrology and EarthSystem Sciences 16: 1445–1463.

Bouvenot, J.-B., M. Siroux, B. Latour, and B. Flament. 2015.“Dwellings Electrical and DHW Load Profiles GeneratorsDevelopment for μCHP Systems Using RES Coupled toBuildings Applications.” Paper presented at the 6th Inter-national Building Physics Conference (IBPC 2015), EnergyProcedia 78 (2015) 1919–1924. doi:10.1016/j.egypro.2015.11.372.

Burhenne, S., D. Wystrcil, M. Elci, and S. Narmsara. 2013.“Building Performance Simulation Using Modelica: Anal-ysis of the Current State and Application Areas.” Proceed-ings of BS2013: 13th Conference of International BuildingPerformance Simulation Association, Chambéry, France,August 26–28.

Carmo, C., O. Dumont, B. Elmegaard, and Detlefsen Nielsen.2015. “Experimental Validation of a Domestic StratifiedHot Water Tank Model in Modelica for Annual Per-formance Assessment.” ORC 2015 Conference, Brussels,Belgium.

CEN. 2008. prEN12977-3. Thermal Solar Systems and Compo-nents – Custom Built Systems – Part 3: Performance TestMethods for Solar Water Heater Stores. Technical report.

Chow, T. 2003. “Performance Analysis of Photovoltaic-thermalCollector by Explicit Dynamic Model.” Solar Energy 75:143–152.

Danish Energy Agency. 2015. Accessed 22 July. https://www.retsinformation.dk/forms/r0710.aspx?id = 152758.

Duffie, J. A., and W. A. Beckham. 2006. Solar Thermal Engineer-ing of Thermal Processes. Wisconsin, Madison: John Wileyand Sons.

Dumont, O., C. Carmo, F. Randaxhe, S. Quoilin, and V. Lemort.2015. “Performance Comparison of Two Types of Tech-nologies Associated with a Positive Energy Building: AReversible Heat Pump/ORC Unit and a Heat Pump Cou-pled with PV Panels.” Proceedings of the ISES Solar WorldCongress (SWC 2015). International Solar Energy Society,Liege, Belgium, pp. 1–6.

Dumont, O., S. Quoilin, and V. Lemort. 2014. “ExperimentalInvestigation of a Scroll Unit Used as a Compressor andas an Expander in a Reversible Heat Pump/ORC Unit.”Proceedings of the refrigeration conference in Purdue.

Dumont, O., S. Quoilin, and V. Lemort. 2015. “ExperimentalInvestigation of a Reversible Heat Pump/Organic Rank-ine Cycle Unit Designed to be Coupled with a PassiveHouse (Net Zero Energy Building).” International Jour-nal of Refrigeration 54: 190–203. http://dx.doi.org/10.1016/j.ijrefrig.2015.03.008.

Dupeyrat, P., P. Ménézo, and S. Fortuin. 2014. “Study of the Ther-mal and Electrical Performances of PVT Solar Hot WaterSystem.” Energy and Buildings 68: 751–755.

Energinet. 2015. Accessed 2 January. http://energinet.dk/DA/El/Solceller/Har-du-solceller/Sider/Pristillaeg.aspx.

EnergyPlus. 2015. http://apps1.eere.energy.gov/buildings/energyplus/weatherdata_about.cfm.

European Commission. 2012. Energy Markets in the EuropeanUnion in 2011. doi:10.2833/12806.

European Commission. 2013. Progress by Member Statestowards Nearly Zero-Energy Buildings, COM (2013) 483.

EV-power. 2014. Accessed 29 October. http://www.ev-power.eu/Winston-300Ah-1000Ah/.

Fischer, S., W. Heidemann, H. Müller-Steinhagen, B. Perers, P.Bergquist, and B. Hellströmd. 2004. “Collector Test Methodunder Quasi-dynamic Conditions According to the EuropeanStandard EN 12975-2.” Solar Energy 76: 117–123.

Freeman, J., K. Hellgardt, and C. Markides. 2015. “An Assess-ment of Solar-Powered Organic Rankine Cycle Systems for

Journal of Building Performance Simulation 13

Combined Heating and Power in UK Domestic Applica-tions.” Applied Energy 138: 605–620.

Georges, E., S. Gendebien, S. Bertagnolio, and V. Lemort. 2013.“Impact of the Integration of Various Heating Technolo-gies on the Energy Load Profiles of the Belgian ResidentialBuilding Stock.” Proceedings of the 8th renewable energystorage conference and exhibition (IRES 2013).

He, W., Tin-Tai Chow, Jie Ji, Jianping Lu, G. Pei, and Lok-shun Chan. 2006. “Hybrid Photovoltaic and Thermal Solar-Collector Designed for Natural Circulation of Water.”Applied Energy 83: 199–210.

Herrando, M., C. Markides, and K. Hellgardt. 2014. “A UK-Based Assessment of Hybrid PV and Solar-Thermal Sys-tems for Domestic Heating and Power: System Perfor-mance.” Applied Energy 122: 288–309.

Innogie ApS. 2013. “Thermal Solar Absorber System Gener-ating Heat and Electricity.” US Patent, 2013/025778 A1,https://patentscope.wipo.int/search/en/detail.jsf?docId =WO2012041323&recNum = 27&docAn = DK2011050363&queryString = (FP/solar%20AND%20FP/thermal%20AND%20FP/electricity)%20&maxRec = 107.

Klein, S. A. 1975. “Calculation of Flat-Plate Loss Coefficient.”Solar Energy 17: 79. http://adsabs.harvard.edu/abs/1975SoEn . . . 17 . . . 79K.

Knight, I., J. Adnot, J.-L. Alexandre, P. Andre, M.-N. Assi-makopoulos, V. Butala, R. Hitchin, M. Masoero, C.Spitzbart, and D. Wright. 2010. “HARMONAC – Har-monizing Air Conditioning Inspection and Audit, Proce-dures in the Tertiary Building Sector, Energy Consump-tion in European Air Conditioning Systems and the Air.”https://ec.europa.eu/energy/intelligent/projects/en/projects/harmonac.

Marszal, A. J., P. Heiselberg, J. S. Bourrelle, E. Musall, K. Voss,I. Sartori, and A. Napolitano. 2011. “Zero Energy Building– A Review of Definitions and Calculation Methodologies.”Energy and Buildings 43 (4): 971–979.

Masy, G. 2007. “Definition and Validation of a Simplified Multi-zone Dynamic Building Model Connected to Heating Sys-tem and HVAC Unit.” PhD thesis, University of Liege,Liege, Belgium.

Masy, G., E. Georges, C. Verhelst, V. Lemort, and P. André.2015. “Smart Grid Energy Flexible Buildings through theuse of Heat Pumps and Building Thermal Mass as EnergyStorage in the Belgian context.” Science and Technologyfor the Built Environment – Ashrae 21 (6): 8000–811.doi:10.1080/23744731.2015.1035590.

Muller, D., and A. H. Badakhshani. 2010. “Gebaude und Anla-gensimulati Mit Modelica.” Proceedings of BauSIM 2010.http://info.tuwien.ac.at/bausim/conftool/mueller-2010-gekoppelte_gebaeude-_und_anlagensimulation_mit_modelica-198.pdf.

Nytsch-Geusen, M., and S. Unger. 2009. “Modelling of Com-plex Thermal Energy Supply Systems Based on the

Modelica-Library Fluidflow.” Proceedings of the 6th Inter-national Modelica Conference. http://webcache.googleusercontent.com/search?q = cache:D2BJ4wVYr8IJ:citeseerx.ist.psu.edu/viewdoc/summary3FDOI3D10.1.1.472.1996 + &cd= 3&hl = fr&ct = clnk&gl = be.

Perers, B. 1997. “An Improved Dynamic Solar Collector TestMethod for Determination of Non-Linear Optical and Ther-mal Characteristics with Multiple Regression.” Solar Energy59: 163–178.

PICIR (Potsdam Institute for Climate Impact Research). 2015.Accessed 30 July. www.pik-potsdam.de.

Quoilin, S., A. Desideri, J. Wronski, and I. Bell. 2014. “Ther-moCycle: A Modelica Library for the Simulation of Ther-modynamic Systems.” Proceedings of the 10th InternationalModelica Conference 2014, Lund, Sweden.

Quoilin, S., O. Dumont, K. Harley, and V. Lemort. 2015. “Design,Modeling and Performance Optimization of a ReversibleHeat Pump/Organic Rankine Cycle System for DomesticApplication.” Journal of Engineering for Gas Turbines andPower 138: 011701. doi:10.1115/1.4031004.

Quoilin, S., O. Dumont, and V. Lemort. 2013. “Design, Mod-elling and Performance Optimization of a ReversibleHP/ORC Prototype.” Proceedings of the ORC 2013 Confer-ence, Rotterdam, Netherlands.

Schimpf, S., K. Uitz, and R. Span. 2011. “Simulation of a SolarAssisted Combined Heat Pump-Organic Rankine Cycle Sys-tem.” Proceedings of World Renewable Energy Congress,Sweden.

Schnieders, J. 1997. “Comparison of the Energy Yield Predic-tions of Stationary and Dynamic Solar Collector Models andthe Models’ Accuracy in the Description of a Vacuum TubeCollector.” Solar Energy 61: 179–190.

TESS. 2006. “Standard Component Library Overview.” http://web.mit.edu/parmstr/Public/Documentation/03-ComponentLibraryOverview.pdf.

Trnsys. 2015. “Tess Component Libraries.” Accessed 1 Septem-ber. http://www.trnsys.com/tess-libraries/TESSLibs17_General_Descriptions.pdf.

Wang, P. G., M. Scharling, K. P. Nielsen, and C. Kern-Hansen.2010. “Technical Report 13–19: 2001 – 2010 Danish DesignReference Year, Climate Dataset for Technical Dimension-ing in Building, Construction and other Sectors.” Accessed1 September 2015. http://www.dmi.dk/fileadmin/Rapporter/TR/tr13-19.pdf.

Wetter, M., W. Zuo, and T. S. Nouidui. 2011. “RecentDevelopments of the Modelica ‘Buildings’. Library forBuilding Energy and Control Systems.” Proceedings ofthe 8th International Modelica Conference, Dresden,Germany.

Wetter, M., W. Zou, T. S. Nouidui, and X. Pang. 2013. “Mod-elica Buildings Library.” Journal of Building PerformanceSimulation, March 2013, doi:10.1080/19401493.2013.765506

14 O. Dumont et al.

AppendixFigure A1 presents the T-s diagram for the HP mode (dashed linesare the theoretical cycle, plain lines are the experiment). The dif-ferent steps of the process are observed: expansion valve exhaustand evaporator supply (1), evaporator exhaust and four-way valve(low pressure) supply (2), four-way valve exhaust and compres-sor supply (3), compressor exhaust and four-way valve (highpressure) supply (4), four-way valve exhaust and condenser sup-ply (5), condenser exhaust and sub-cooler supply (6), sub-coolerexhaust and expansion valve supply (7). (Dumont, Quoilin, andLemort 2015). More details are provided in Dumont, Quoilin, andLemort (2015).

Figure A2 depicts the T-s diagram for the ORC mode (com-parison between experiments and theory). The different steps ofthe process are observed: Sub-cooler exhaust and pump supply(1), pump exhaust and evaporator supply (2), evaporator exhaustand four-way valve supply (high pressure) (3), four-way valveexhaust and expander supply (4), expander exhaust and four-wayvalve (low pressure) supply (5), four-way exhaust and condensersupply (6) and, condenser exhaust and sub-cooler supply (7).(Dumont, Quoilin, and Lemort 2015). More details are providedin Dumont, Quoilin, and Lemort (2015).

Figure A3 shows the hydraulic scheme of the thermal heatstorage. Unit loop (supply and return) is the reversible HP/ORCunit (direct heating or heat pump). In heat pump mode, it is con-nected to the condenser and in direct heating mode is connectedto the solar roof. FH is the floor heating loop while DHW is theDomestic Hot water loop.

Figure A1. T-s diagram for the HP mode (dashed lines are thetheoretical cycle, plain lines are the experiment).

Figure A2. T-s diagram for the ORC mode (comparisonbetween experiments and theory). Dashed lines are the theory,plain line is the experiment.

Figure A3. Hydraulic scheme of the thermal heat storage. Unitloop (supply and return) is the reversible HP/ORC unit (directheating or heat pump). In heat pump mode, it is connected to thecondenser and in direct heating mode is connected to the solarroof. FH is the floor heating loop while DHW is the domestic hotwater loop.

Journal of Building Performance Simulation 15

Figure A4. Patented solar roof collector scheme.

Table A1. Meaning and value of the terms of Equation (3).

Term Name Value/expression

β Collector tilt 5 (°C)v Empirical constant 520(1 − 0.000051β2)

e Empirical variable 0.43(

1 − 100Tm

)

εg Emittance of glass 0.88εp Emittance of plate 0.95n Empirical constant (1 + 0.089 · hw −

0.1166hwεp )(1 +0.07866N )

hw Wind heat transfercoefficient

2 (W m−2 K−1)

s Empirical constant 0.00591z Empirical constant 0.133

Figure A4 presents the patented solar roof collector scheme.This roof acts is an integrated solution replacing a classical roof.The pipes are composed of aluminium.

Figure A5 presents the layout of the GHX. The right-handloop is a cooling system and the bottom loops are the GHX withtwo main hoses of connection.

16 O. Dumont et al.

Table A2. Five zones of the house characteristics. For each zone, the floor area, the total volume, the heat exchange coefficient(U-value) for the roof, the slab, the wall and the window are given. The area of windows and walls, the infiltration rate, the nominallighting and appliances power and the air–temperature set-points are also provided.

Unit Zone 1 Zone 2 Zone 3 Zone 4 Zone 5

Floor area m2 41.8 18.2 7.8 19.1 45.7Volume m3 117.2 45.5 19.5 47.8 114.3Slab U-Value W/m2 K 0.08 0.08 0.08 0.08 0.08Roof U-Value W/m2.K 0.09 0.09 0.09 0.09 0.09External wall area m2 none 20.4 4.5 24.8 41.5External wall U-value W/m2.K none 0.15 0.15 0.15 0.15Window area (orientation) m2 14.7(S) 2.4(S) 0.84(W) 0.84(W) 6.7(E)

0.84(N) 2.4(S)Window U-value W/m2 K 0.63 0.68 0.8 0.8 0.8Window solar factor – 0.5 0.5 0.5 0.5 0.5Infiltration rate ACH 0.3 0.3 0.3 0.3 0.3

Space activity –KitchenDining Main Bedroom Bathroom Hall Others Living Bedroom

Lighting nominal power W/m2 5 5 3 3 5Appliances nominal power W/m2 3 3 3 3 3

Air–temperature Set-point °C 20

Onlyimposed in

zone 1

Onlyimposed in

zone 1

Onlyimposed in

zone 1

Onlyimposed in

zone 1

Figure A5. Layout of the GHX. The right-hand loop is a cooling system and the bottom loops are the GHX with two main hoses ofconnection.

Journal of Building Performance Simulation 17

Table A3. Parameters of the reduced-order model calibratedbased on the finite-element model response. The heat exchangebetween the pipes and the central soil is modelled with threeresistances in series (soil, tube and convective). The thermalbehaviours of the surface and sub-soil are modelled through athermal capacity and a thermal resistance. An additional convec-tive resistance is added to the surface layer to take into accountthe exchange with the air. The related thicknesses allow to eval-uate the mass of each layer. The inertia of the central soil ischaracterized by an equivalent thermal capacity.

Parameter Value Unit

Surface Thermal capacity 3E07 J/KRelated thickness 0.033 mThermal resistance 0.0011 K/WConvective resistance 1.58E − 04 K/W

Sub-soil Thermal capacity 4E09 J/KRelated thickness 4.47 mThermal resistance 0.005 K/W

Central soil Thermal capacity 1.2E09 J/KRelated thickness 2.01 m

Pipes Convective resistance 1.26E − 04 K/WTube resistance 4.89E − 05 K/WSoil resistance 2E − 04 K/W

Contractsurface

Area 299 m2

Table A4. Envelope characteristics of different typicalbuildings. Each building is characterized by different coef-ficients of heat transmission for the roof, the floor slab,the external wall and the window and an infiltration ratecoefficient.

Coefficient of heattransmission

Danish K15 K30

Roof (W m−2 K−1) 0.09 0.093 0.228Floor slab (W m−2 K−1) 0.08 0.123 0.258External wall

(W m−2 K−1)0.15 0.102 0.245

Window (W m−2 K−1) 0.63 0.9 1.2Infiltration rate (50 Pa)

(m3 h−1 m−2)2.51 0.6 0.35